diff --git a/genkit-tools/cli/src/commands/eval-flow.ts b/genkit-tools/cli/src/commands/eval-flow.ts
index cbfb3fb2d..3245fca41 100644
--- a/genkit-tools/cli/src/commands/eval-flow.ts
+++ b/genkit-tools/cli/src/commands/eval-flow.ts
@@ -322,7 +322,7 @@ async function fetchDataSet(
         testCaseId: randomUUID(),
         input: inputs[0],
         output: outputs[0],
-        context: contexts,
+        context: JSON.parse(contexts[0]) as string[],
         reference: references?.at(i),
         traceIds,
       };
diff --git a/js/testapps/cat-eval/src/index.ts b/js/testapps/cat-eval/src/index.ts
index 5ac4f9ccd..b3a6784aa 100644
--- a/js/testapps/cat-eval/src/index.ts
+++ b/js/testapps/cat-eval/src/index.ts
@@ -19,7 +19,7 @@ import { devLocalVectorstore } from '@genkit-ai/dev-local-vectorstore';
 import { dotprompt } from '@genkit-ai/dotprompt';
 import { genkitEval, GenkitMetric } from '@genkit-ai/evaluator';
 import { firebase } from '@genkit-ai/firebase';
-import { geminiPro, googleAI } from '@genkit-ai/googleai';
+import { gemini15Pro, googleAI } from '@genkit-ai/googleai';
 import { textEmbeddingGecko, vertexAI } from '@genkit-ai/vertexai';
 
 // Turn off safety checks for evaluation so that the LLM as an evaluator can
@@ -51,7 +51,7 @@ configureGenkit({
     firebase(),
     googleAI(),
     genkitEval({
-      judge: geminiPro,
+      judge: gemini15Pro,
       judgeConfig: PERMISSIVE_SAFETY_SETTINGS,
       metrics: [GenkitMetric.MALICIOUSNESS],
       embedder: textEmbeddingGecko,